Quote Originally Posted by gabe View Post
Think travel distance. Screws bottoms out already. It doesn't matter how much spring tension. The waterblock base can't travel down to mate with IHS.
I understand the screw bottoms out, I was thinking more along the lines of increasing spring tension. During testing of the GTZ on a lapped processor, I never noticed where there was some sort of travel distance limitation, there was always quite a bit of spring compression and well within the travel limits of the system on a lapped processor. Of coarse I don't lap my processors very much, only enough to remove the coating and level the surface over the core, so it's probably a very small thickness. There have also been some past work by Bill Adams showing somewhat of a progression in performance relative to mounting pressure. I haven't tried plotting this out myself, but it seems reasonable to presume mounting pressure does have an important role. Not sure more is always better, but it does play a role depending on the thermal compound used. If the thought was that lapping a processor causes a loss of thickness and that's the only problem here, this should be correctable with the use of an additional washer between the retaining plate and spring. If you lost .5mm of thickness from lapping (I'm sure it's not that much) a .5mm washer should correct that and apply the same spring compression, etc. At least that's the idea I had.

Regardless, I've always wondered if there was any difference. My thought on lapping in the past was to ensure a flat contact surface since it appears the actual IHS shape may vary to some degree between samples. If the IHS shape is indeed sample specific, then you could possibly see block results vary per sample. A concaved IHS for example would likely see more favorable gains from a highly bowed base. Or a convex IHS may see little to no improvement in bowing. My thought was, lapping the base helps eliminate that sample variable to some degree. This is all speculation though, as hobby testers we simply don't have the ability to buy and test half a dozen processors of the same type to do that sort of analysis and I haven't really seen much of any work in that area. I haven't even seen too much more than very casual testing of stock vs. lapped IHS testing.

Anyhow, there could be a difference, but I would be suspect of IHS sample specific differences as well. I know with the few I've lapped, they all seem to have some variation, so trying to test and verify that may take retesting over several samples..not sure.

Quote Originally Posted by gabe View Post
I do not intend to keep it to myself. You know me better than that
I am working with an independent tester. The strategy is that if the data is presented by a well-known independent, it will be accepted more readily. Too many people (even if they do not openly say so) question the data I present. just read these comments: http://www.xtremesystems.org/forums/...3&postcount=53 , it's fairly typical, and very frustrating to me.

So I will let him present his data, and then be available for comments and additional input.

I just wish for the community to accept the findings as unbiased, that's all.
I don't know how many times I've compared something like a pressure drop curve and been pretty well spot on with your results. I have no reason not to look forward to it, and have always appreciated the information.